4512 YASUHARA AND NOGI
The Journal of Organic Chemistry
Carboxylation by Alkali Salts and Carbon Monoxide. 11.' A Selective Preparation of p-Hydroxybenzoic Acid YUTAEAYASUHARA AND TATSUO NOGI Mishima Laboratory, Toyo Rayon Company, Ltd., Mishimu, Japan Received March 8, 1968 The reaction of potassium phenoxide with carbon monoxide and potassium carbonate was studied in detail. The effects of reaction variables on the yield and the selectivity were investigated. The reaction proceeds very smoothly at a temperature above the melting point of potassium formate (167") even under low pressure of carbon monoxide to afford dipotassium phydroxybenzoate and potassium formate in very high yield. The carboxylation occurs mainly a t the para position to the hydroxyl group. When the molar ratio of potassium carbonate to potassium phenoxide is larger than unity, tripotassium 4-hydroxyisophthalate is also formed by the subsequent but rather slow carboxylation.
p-Hydroxybenzoic acid (or its potassium salt) has been prepared by various procedures. One of the most practical is the Kolbe-Schmitt r e a ~ t i o n . ~ ~ ~ On the other hand, the carboxylation of potassium benzoate by carbon monoxide and potassium carbonate to give dipotassium terephthalate was reported by Murase, et aL4 Although the mechanism of the reaction is not clear, the reaction probably proceeds through the following steps: potassium carbonate reacts first with carbon monoxide to form potassium oxalate5 as an intermediate, and then the potassium oxalate converts potassium benzoate into dipotassium terephthalak6 The carboxylation reaction of potassium phenoxide with carbon monoxide and potassium carbonate was reported in the previous paper.' The reaction is investigated in detail in this paper. Results and Discussion Effects of the Reaction Variables.-The equimolar mixture of potassium phenoxide and potassium carbonate was heated under fairly high pressure of carbon monoxide at an appointed temperature for 5 hr (Figure 1). Even at, 125", the carboxylated products, though the conversions were very low, were detected. The yield and the selectivity of p-hydroxybenzoic acid increased slowly with temperatures up 170". At above 170°, near the melting point of potassium formate (167"), an abrupt increase of the yield was observed. At temperatures above 210°, the yield of p-hydroxybenzoic acid was almost quantitative. The almost stoichiometrical absorption of carbon monoxide and the almost quantitative formation of formic acid show the reaction to be formulated as
8
-0 OK
OK
+.K2C03
+
CO
I
+
liquid phase in the reaction mixture. At temperatures above 210°, the change of carbon monoxide pressure indicated that the reaction rate depends considerably on the temperature (Figure 2). The effect of the carbon monoxide initial pressure on the yield of p-hydroxybenzoic acid was examined a t 240" (Figure 3). When the reaction was carried out in the absence of carbon monoxide, no carboxylated product was obtained. When carbon monoxide was present enough for the amount of potassium phenoxide (0.050 mol of carbon monoxide corresponds to a pressure of about 15 atm, at room temperature in our system), the yield of p-hydroxybenzoic acid was nearly quantitative irrespective of the carbon monoxide initial pressure. This shows that the reaction proceeds even under low pressure. Figure 4 indicates clearly that the reaction rate is affected by the initial pressure. The effect of the molar ratio of potassium carbonate to potassium phenoxide was studied at 240" (Figure 5). The yield of p-hydroxybenzoic acid corresponded to the proportion of potassium carbonate up to the ratio of unity. When 1.5 or 2.0 times as much potassium carbonate was used, a small amount of 4-hydroxyiso-
OK
OK
I
I
CO2K
C02K
phthalic acid was obtained. The formation of it is by the subsequent carboxylation of dipotassium p-hydroxybenzoate, but the reaction rate was so slow that usually p-hydroxybenzoic acid can be obtained preferentially (Table I).
HCOZK (1)
bOzK
One of the reasons of the low reaction rate a t a temperature below 167' is presumably that there is no (1) Y. Yasuhara and T. Nogi, Chem. Ind. (London), 229 (1967),may be considered as part I in this series. (2) H. Kolbe, J . Prakt. Chem., [21 10, 100 (1874). (3) A. S. Lindsey and H. Jeskey, Chem. Rev.,67, 583 (1957). (4) T . Murase, I. Mikami, and M. Tamura, British Patent 1,003,725 (1965);Chem. Abstr., 64,655b (1966). (5) C. Matignon and C. Faurholt, C . R . Acad. Sei., Paris, 179, 273 (1923). (6) W. Stein and H. Schutt, German Patent 1,042,568 (1958); Chem. Abstr., 64, 184440 (1960).
TABLE I CARBOXYLATION OF DIPOTASSIUM p-HYDROXYBENZOATE~ Solvent, H.COpK, mol
Initial CO, atm
Time, hr
-Yield of product? %4-HIPA HTMA p-HBA
0 50 5.0 0 0 97 0.050 50 5.0 14 0 84 0.050 50 24.0 43 4 50 0.050 10 5.0 0 0 97 0.050 10 24.0 3 0 95 a A mixture of 0.050 mol of dipotassium p-hydroxybenzoate and 0.050 mol of potassium carbonate was heated a t 240". b 4-HIPA, 4-hydroxyisophthalic acid; HTMA, hydroxytrimesic acid; p-HBA, p-hydroxybenzoic acid.
A SELECTIVE PREPARATION OF p-HYDROXYBENZOIC ACID 4513
Vol. $8, No. 1.2, December 1968 100
t3
80
a
s
20 0 120
140
160 180 200 Temperature, OC
220
240
6'
c)
u
0
Figure 1.-Over-all yield of hydroxybenzoic acids (-0-) and selectivity of the para isomer (- -0--) as a function of temperature. A mixture of 0.050 mol of PhOK and 0.050 mol of K&Os was heated under the initial pressure (80 atm) of CO a t an a p pointed temperature for 5 hr.
160 140
A
0
A
A
20 40 60 80 Initial pressure of CO, atm.
100
Figure 3.-Yield of p-hydroxybenzoic acid (--e) and recovery of PhOH (- -0--) as a function of initial pressure of CO. A mixture of 0.050 mol of PhOK and 0.050 mol of K&Oa was heated under an appointed initial pressure of CO at 240' for 3 hr.
r a a
i
160 120
*
;
120
f
2 100 &
0 1 .
' 0
I 1 Time, hr.
I
2
I 3
1
Figure 4.-Reaction velocity expressed by change of CO pres20 atm; sure as a function of the initial CO pressure (-*, -0--,40 atm; -@- * -, 80 atrn). For the reaction conditions, see Figure 3.
-
-
The fact that 4-hydroxyisophthalic acid was not detected in the reaction of dipotassium p-hydroxybenzoate with carbon monoxide in the absence of potassium formate may be understood in terms of the absence of liquid phase in the reaction mixture. Mechanism of the Carboxylation Reaction.-To elucidate the course of reaction 1, the behavior of dipotassium salicylate under the present reaction conditions was examined (Table 11). (I
TABLE I1 CARBOXYLATION O F DIPOTASSIUM SALICYLATEo Solvent, HCOzK, mol
ob
Temp, 'C
Time, hr
-Yield SA
of produot,d %p-HBA 4-HIPA
175 5.0 95 0 3 240 0.4 4 90 2 240 5.0 0 93 5* 0.025" 170 5.0 86 5 7 170 20.0 57 25 16 240 0.4 0 93 2 240 5.0 0 92 4 a The reagents were 0.050 mol of dipotassium salicylate and 0.050 mol of potassium carbonate. b The initial pressure of CO The initial pressure of CO was 60 atm. was 80 atm. SA, salicylic acid; p-HBA, p-hydroxybenzoic acid ; 4-HIPA, 4-hydroxyisophthalic acid. e Mainly hydroxytrhesic acid.
0.5 1.0 1.5 Molar ratio KzCOa/PhOK.
2.0
Figure 5.-Yield of p-hydroxybenzoic acid (d) and 4hydroxyisophthalic acid (- -0- -) and recovery of PhOH (- -@. -) &s a function of molar ratio of KzCOI to PhOK. A mixture of 0.040 mol of PhOK and an appointed amount of &COS W&S heated under the initial pressure (50 atm) of CO a t 240' for 5 hr. a
At temperatures below 175", the yield of p-hydroxybenzoic acid obtained by the rearrangement of dipotassium salicylate was much lower than that by the carboxylation of potassium phenoxide, irrespective of the presence of potassium formate. These results lead to the conclusion that most of p-hydroxybenzoic acid was formed by the direct para carboxylation, a t least at temperature below 175".
4514
YASUHARA AND
NOGI
At 240", where dipotassium salicylate rearranges rapidly to the para isomer in the Kolbe-Schmitt rethe rapid rearrangement was also observed accompanied with small extent of the polycarboxylation. Since the rearrangement is inhibited at the same temperature in the presence of 1 mol of potassium hydroxidelO+ll(or methoxide**lo)in the Kolbe-Schmitt reaction, the carboxylation was examined also in the presence of potassium hydroxide. The potassium hydroxide, however, reacted smoothly with carbon monoxide to give potassium formate. l 2 ~ l 3 Therefore, at 240" a possibility of the initial ortho carboxylation was not thoroughly avoided. The assumption that potassium oxalate is an intermediate carboxylating agent of potassium phenoxide should be abandoned, because both were recovered on heating them together at 250' for 6 hr. Experimental Section14 Reagents.-Potassium phenoxide, dipotassium p-hydroxybenzoate, and potassium formate were prepared by neutralization of aqueous n-propyl alcohol or aqueous n-butyl alcohol solutions, followed by distillation of solvents to yield crystals, and drying the crystals at 130' in vacuo (Nz, 20 mm). Dipotassium salicylate was difficult to crystallize, and so was prepared by drying an aqueous methanol solution a t 150' in vacuo. The absence of p-hydroxybenzoic acid which may be formed by the thermal rearrangement on drying was confirmed by extraction of the regenerated acid with chloroform.16 High purities (97-98%) were proved by potentiometric titration (and gravimetry of the regenerated acid) in all cases. Potassium oxalate was used after drying the commercial monohydrate a t 160" in VQCUO. Potassium carbonate and carbon monoxide (Matheson Co.) were used without further purification. Procedure .-The pulverized reagents (usually 0.050 mol of potassium phenoxide and 0.050 mol of potassium carbonate) were mixed well, and then put into a glass vessel having a coiled capillary vent together with a steel ball. The vessel, after evacuation (Nz,2-20 mm) a t 150' for 5 hr, was placed in an autoclave (ca. 100 ml). The autoclave was filled with carbon monoxide to an appointed pressure at room temperature. Then the autoclave was settled in an aluminum block heater regulated within f 2 ' and agitated vigorously. About 20 min were required to bring the temperature of the autoclave constant. After heating for an appointed time, the autoclave was withdrawn and the gas pressure was recorded. The pressure drop reflecting usually the extent of the carboxylation. The gas was submitted to gas chromatography [column of activated charcoal (2 m, go'), carrier gas, He (50 ml/min)] and proved usually to (7) A. J. Rostron and A. M. Spivey, J . Chem. Soc., 3093 (1964). (8) S. E. Hunt, J. I. Jones, A. 9. Lindsey, D . C. Killoh, and H. 9. Turner, ibid., 3152 (1958). (9) C. A. Buehler and W. E . Cate, "Organic Syntheses," Call. Vol. 11, John Wiley & Sons, Inc., New York, N . Y., 1943, p 341. (10) J. I. Jones, A. S. Lindsey, and H. S. Turner, Chem. Ind. (London), 659 (1958). (11) H. Ost, J . Prakt. Chem., [2] 11, 385 (1875). (12) M. Berthelot, Ann., 97, 125 (1856). (13) A. Gleuther and 0. Froelich. ibid., 000, 288 (1880). (14) All melting points mere determined by the capillary method and were
corrected. (15) H. Kolbe, .I. Prakt. Chem., [2] 10, 120 (1874).
The Journal of Organic Chemistry consist of carbon monoxide with only a trace amount of carbon dioxide. The glass vessel was weighed and the increase of the weight was generally equal to the amount of absorbed carbon monoxide. The crude product under the typical reaction conditions (0.050 mol of PhOK, 0.050 mol of K2C03,50 atm of GO, 240", 5 hr) showed practically the same ir spectra and also the same titration curve as an equimolar mixture of dipotassium p-hydroxybenzoate and potassium formate. Isolation and Analysis of Product .-The reaction mixture was dissolved in 50 ml of water and neutralized to pH 9.0 with 6 N sulfuric acid, and the regenerated phenol was extracted with ether. The solid phenol (identified by the ir spectra), obtained after evaporation of the solvent and drying, was weighed. After acidification with 12 N sulfuric acid to pH 1.0, the aqueous solution was extracted again with ether. The crude p-hydroxybenzoic acid, obtained from the ether extract, was weighed, and its melting point and ir spectra were determined. The crude acid usually melted at 208-212', and showed the practically identical ir spectra with that of the pure acid. After recrystallization of the crude acid from water and drying in vacuo (N2, 10 mm) at 130' for 8 hr, p-hydroxybenzoic acid melted a t 213-214' (lit.16mp 213-214'). Anal. Calcd for C7HeO3: C, 60.86; H, 4.38. Found: C, 60.92; H , 4.45. Separation of the other acid from the crude p-hydroxybenzoic acid, if necessary, was carried out as follows. Salicylic acid was extracted with 160 ml of chloroform16from 50 ml of an aqueous suspension of the crude acid, and 4-hydroxyisophthalic acid was separated from the aqueous solution of p-hydroxybenzoic acid by fractional recrystallization.17~1~Salicylic acid and 4-hydroxyisophthalic acid thus obtained were identified by the mixture melting points with authentic samples and by the ir spectra. I n certain cases (for example, dipotassium p-hydroxybenzoate, 0.050 mol; KzCOs, 0.050 mol; HCOtK, 0.050 mol; CO, 50 atm, 240°, 24 hr), hydroxytrimesic acid, mp 304-306' (crystallized from water, lit.16 mp 306'), was isolated together with 4hydroxyisophthalic acid. Anal. Calcd for COHnO,: C. 47.80: H, 2.67. Found: C, . . 48.23; H, 2.81. I n some cases, formic acid was titrated potentiometrically in the following manner. The ttqueous solution of the reaction product, after removal of the regenerated phenol, was acidified to pH 1.0 and extracted exhaustively with ether. The extract was titrated with 0.5 N sodium hydroxide in aqueous ethanol. The yield of formic acid was calculated from the difference between the total amount of formic acid and p-hydroxybenzoic acid which was determined from the end point of the titration, and the amount of p-hydroxybenzoic acid weighed by the above procedure. Generally the yield of formic acid was equal to that of p-hydroxybenzoic acid. Formic acid (bp 100-101") was isolated from the aqueous solution, from which p-hydroxybenzoic acid had been removed, by exhaustive extraction with ether and fractional distillation; the ir spectra was consistent with that of an authentic sample; the anilide10 had mp 46-48' (lit." mp 50'). Anal. Calcd for C7H7NO: C, 69.40; H, 5.83; N , 11.57. Found: C, 69.25; H, 5.89; N, 11.62.
Registry No.-Carbon monoxide, 630-08-0; potassium carbonate, 584-08-7; potassium phenoxide, 100-67-4 ; p-hydroxybenzoic acid, 99-96-7. (16) I. Heilbron, "Dictionary of Organic Compounds," Vol. 3, Eyre & Spottiswoode Ltd., London, 1965, p 1652. (17) J. Walker and J. K. Wood, J . Chem. Soc.. 73, 622 (1898). (18) H. Ost, J . PrakZ. Chem., [2] 14, 103 (1876). (19) G. Tobias, Ber., 16, 2443 (1882). (20) See ref 16, p 1456.
